EDA, FE and Regression Models (Car Dekho Dataset)

1. EDA and FE

1. Data Profiling
2. Stastical analysis
3. Graphical Analysis
4. Data Cleaning
5. Data Scaling

2. Linear Regression Model

1. Linear Regression Model
2. Performance metrics for above model

3. Ridge, Lasso and Elastic-Net Regression Model

1. Ridge, Lasso and Elastic-Net Regression Model
2. Performance metrics for above model

4. Decision Tree Regressor Model

1. Decision Tree Regressor Model
2. Performance metrics for above model
3. Hyper-Parameter Tuning for above model

5. Support Vector Regressor Model

1. Support Vector Regressor Model
2. Performance metrics for above model
3. Hyper-Parameter Tuning for above model

6. Random Forest Regressor Model

1. Random Forest Regressor Model
2. Performance metrics for above model
3. Hyper-Parameter Tuning for above model

Dataset: https://github.com/Shubham3023/Machine_Learning/blob/main/Linear%20Reg%2C%20SVR%2C%20DTR%20and%20RFR%20for%20CarDekho%20dataset/cardekho_dataset.csv

from IPython import display
display.Image("image.jpg")

1.0 Importing required libraries

### Pandas and Numpy
import pandas as pd
import numpy as np

### Visualisation libraries
import seaborn as sns
import matplotlib.pyplot as plt
%matplotlib inline

### For Q-Q Plot
import scipy.stats as stats

### To ignore warnings
import warnings
warnings.filterwarnings('ignore')

### Machine Learning libraries
import sklearn
from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.preprocessing import StandardScaler
from sklearn.svm import SVR
from sklearn.tree import DecisionTreeRegressor
from sklearn.ensemble import RandomForestRegressor
from sklearn.linear_model import LinearRegression, Lasso, Ridge, ElasticNet
from sklearn.metrics import r2_score

### To be able to see maximum columns on screen
pd.set_option('display.max_columns', 500)

2.0 Importing dataset and performing descriptive analysis

dataset=pd.read_csv('cardekho_dataset.csv', usecols=['car_name', 'brand', 'model', 'vehicle_age', 'km_driven',
       'seller_type', 'fuel_type', 'transmission_type', 'mileage', 'engine',
       'max_power', 'seats', 'selling_price'])
dataset.head()

	car_name	brand	model	vehicle_age	km_driven	seller_type	fuel_type	transmission_type	mileage	engine	max_power	seats	selling_price
0	Maruti Alto	Maruti	Alto	9	120000	Individual	Petrol	Manual	19.70	796	46.30	5	120000
1	Hyundai Grand	Hyundai	Grand	5	20000	Individual	Petrol	Manual	18.90	1197	82.00	5	550000
2	Hyundai i20	Hyundai	i20	11	60000	Individual	Petrol	Manual	17.00	1197	80.00	5	215000
3	Maruti Alto	Maruti	Alto	9	37000	Individual	Petrol	Manual	20.92	998	67.10	5	226000
4	Ford Ecosport	Ford	Ecosport	6	30000	Dealer	Diesel	Manual	22.77	1498	98.59	5	570000

Feature Information

1. car_name: Car's Full name, which includes brand and specific model name.
2. brand: Brand Name of the particular car.
3. model: Exact model name of the car of a particular brand.
4. seller_type: Which Type of seller is selling the used car
5. fuel_type: Fuel used in the used car, which was put up on sale.
6. transmission_type: Transmission used in the used car, which was put on sale.
7. vehicle_age: The count of years since car was bought.
8. mileage: It is the number of kilometer the car runs per litre.
9. engine: It is the engine capacity in cc(cubic centimeters)
10. max_power: Max power it produces in BHP.
11. seats: Total number of seats in car.
12. selling_price: The sale price which was put up on website.

# getting null values and datatypes of all features
dataset.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 15411 entries, 0 to 15410
Data columns (total 13 columns):
 #   Column             Non-Null Count  Dtype  
---  ------             --------------  -----  
 0   car_name           15411 non-null  object 
 1   brand              15411 non-null  object 
 2   model              15411 non-null  object 
 3   vehicle_age        15411 non-null  int64  
 4   km_driven          15411 non-null  int64  
 5   seller_type        15411 non-null  object 
 6   fuel_type          15411 non-null  object 
 7   transmission_type  15411 non-null  object 
 8   mileage            15411 non-null  float64
 9   engine             15411 non-null  int64  
 10  max_power          15411 non-null  float64
 11  seats              15411 non-null  int64  
 12  selling_price      15411 non-null  int64  
dtypes: float64(2), int64(5), object(6)
memory usage: 1.5+ MB

### getting info about features
dataset.describe(include='all').T

	count	unique	top	freq	mean	std	min	25%	50%	75%	max
car_name	15411	121	Hyundai i20	906	NaN	NaN	NaN	NaN	NaN	NaN	NaN
brand	15411	32	Maruti	4992	NaN	NaN	NaN	NaN	NaN	NaN	NaN
model	15411	120	i20	906	NaN	NaN	NaN	NaN	NaN	NaN	NaN
vehicle_age	15411.0	NaN	NaN	NaN	6.036338	3.013291	0.0	4.0	6.0	8.0	29.0
km_driven	15411.0	NaN	NaN	NaN	55616.480631	51618.548422	100.0	30000.0	50000.0	70000.0	3800000.0
seller_type	15411	3	Dealer	9539	NaN	NaN	NaN	NaN	NaN	NaN	NaN
fuel_type	15411	5	Petrol	7643	NaN	NaN	NaN	NaN	NaN	NaN	NaN
transmission_type	15411	2	Manual	12225	NaN	NaN	NaN	NaN	NaN	NaN	NaN
mileage	15411.0	NaN	NaN	NaN	19.701151	4.171265	4.0	17.0	19.67	22.7	33.54
engine	15411.0	NaN	NaN	NaN	1486.057751	521.106696	793.0	1197.0	1248.0	1582.0	6592.0
max_power	15411.0	NaN	NaN	NaN	100.588254	42.972979	38.4	74.0	88.5	117.3	626.0
seats	15411.0	NaN	NaN	NaN	5.325482	0.807628	0.0	5.0	5.0	5.0	9.0
selling_price	15411.0	NaN	NaN	NaN	774971.11641	894128.363263	40000.0	385000.0	556000.0	825000.0	39500000.0

### checking duplicate values
dataset.duplicated().sum()

167

### dropping duplicate values
dataset.drop_duplicates(inplace=True)

### checking duplicate values after dropping duplicates
dataset.duplicated().sum()

0

3.0 Getting Categorical and Numerical features

### getting count of features
categorical_features=[feature for feature in dataset.columns if dataset[feature].dtypes == 'O']
print("There are {} categorical features namely: {}\n".format(len(categorical_features),categorical_features))

numerical_features=[feature for feature in dataset.columns if dataset[feature].dtypes != 'O']
print("There are {} numerical features namely: {}".format(len(numerical_features),numerical_features))

There are 6 categorical features namely: ['car_name', 'brand', 'model', 'seller_type', 'fuel_type', 'transmission_type']

There are 7 numerical features namely: ['vehicle_age', 'km_driven', 'mileage', 'engine', 'max_power', 'seats', 'selling_price']

### Getting count of categories in each features
for feature in categorical_features:
    print("The {} feature has '{}' no. of different categories".format(feature, dataset[feature].nunique()))

The car_name feature has '121' no. of different categories
The brand feature has '32' no. of different categories
The model feature has '120' no. of different categories
The seller_type feature has '3' no. of different categories
The fuel_type feature has '5' no. of different categories
The transmission_type feature has '2' no. of different categories

### top 10 categories in each feature percent wise
for feature in categorical_features:
    print(dataset[feature].value_counts(normalize=True)[:10]*100)
    print("*******************************************")

Hyundai i20           5.890842
Maruti Swift Dzire    5.739963
Maruti Swift          5.077408
Maruti Alto           5.038048
Honda City            4.919969
Maruti Wagon R        4.651010
Hyundai Grand         3.732616
Toyota Innova         3.568617
Hyundai Verna         3.201260
Hyundai i10           2.663343
Name: car_name, dtype: float64
*******************************************
Maruti        32.360273
Hyundai       19.364996
Honda          9.682498
Mahindra       6.553398
Toyota         5.175807
Ford           5.090527
Volkswagen     4.027814
Renault        3.457098
BMW            2.860142
Tata           2.761742
Name: brand, dtype: float64
*******************************************
i20            5.890842
Swift Dzire    5.739963
Swift          5.077408
Alto           5.038048
City           4.919969
Wagon R        4.651010
Grand          3.732616
Innova         3.568617
Verna          3.201260
i10            2.663343
Name: model, dtype: float64
*******************************************
Dealer              62.050643
Individual          36.814484
Trustmark Dealer     1.134873
Name: seller_type, dtype: float64
*******************************************
Petrol      49.560483
Diesel      48.163212
CNG          1.961427
LPG          0.288638
Electric     0.026240
Name: fuel_type, dtype: float64
*******************************************
Manual       79.336132
Automatic    20.663868
Name: transmission_type, dtype: float64
*******************************************

3.1 Brands and Cars wrt Sales

### Brands wrt car count
pd.DataFrame(dataset.groupby('brand')['brand'].count().sort_values(ascending=False)).rename(columns={'brand':'Car Count'})

	Car Count
brand
Maruti	4933
Hyundai	2952
Honda	1476
Mahindra	999
Toyota	789
Ford	776
Volkswagen	614
Renault	527
BMW	436
Tata	421
Skoda	333
Mercedes-Benz	332
Audi	190
Datsun	170
Jaguar	58
Land Rover	50
Jeep	41
Kia	32
Porsche	21
Volvo	20
MG	19
Mini	17
Nissan	11
Lexus	9
Isuzu	7
Bentley	3
Maserati	2
ISUZU	2
Mercedes-AMG	1
Rolls-Royce	1
Force	1
Ferrari	1

### Top 10 brands wrt sales
plt.figure(figsize=(10,6))
sns.barplot(y=dataset['brand'].value_counts()[:10], x=dataset['brand'].value_counts().index[:10])
plt.ylabel('Count of Sales', fontsize=15, fontweight='bold')
plt.xlabel("Brand",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Top 10 brands wrt Sales",fontsize=15, fontweight='bold')
plt.show();

### Bottom 10 brands wrt sales
plt.figure(figsize=(10,6))
sns.barplot(y=dataset['brand'].value_counts()[-10:], x=dataset['brand'].value_counts().index[-10:])
plt.ylabel('Count of Sales', fontsize=15, fontweight='bold')
plt.xlabel("Brand",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Bottom 10 brands wrt Sales",fontsize=15, fontweight='bold')
plt.show();

### Top 10 Cars
pd.DataFrame(dataset.groupby('car_name')['car_name'].count().sort_values(ascending=False)[:10]).rename(columns={'car_name':'Car Count'})

	Car Count
car_name
Hyundai i20	898
Maruti Swift Dzire	875
Maruti Swift	774
Maruti Alto	768
Honda City	750
Maruti Wagon R	709
Hyundai Grand	569
Toyota Innova	544
Hyundai Verna	488
Hyundai i10	406

### Bottom 10 Cars
pd.DataFrame(dataset.groupby('car_name')['car_name'].count().sort_values(ascending=False)[-10:]).rename(columns={'car_name':'Car Count'})

	Car Count
car_name
ISUZU MUX	2
Isuzu MUX	1
Hyundai Aura	1
Mercedes-AMG C	1
Maserati Quattroporte	1
Maserati Ghibli	1
Rolls-Royce Ghost	1
Ferrari GTC4Lusso	1
Force Gurkha	1
Tata Altroz	1

### Top 10 selling cars countwise
plt.figure(figsize=(10,6))
sns.barplot(y=dataset['car_name'].value_counts()[:10], x=dataset['car_name'].value_counts().index[:10])
plt.ylabel('Count of Sales', fontsize=15, fontweight='bold')
plt.xlabel("Car",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Top 10 selling cars countwise",fontsize=15, fontweight='bold')
plt.show();

### Bottom 10 selling cars countwise
plt.figure(figsize=(10,6))
sns.barplot(y=dataset['car_name'].value_counts()[-10:], x=dataset['car_name'].value_counts().index[-10:])
plt.ylabel('Count of Sales', fontsize=15, fontweight='bold')
plt.xlabel("Car",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Bottom 10 selling cars countwise",fontsize=15, fontweight='bold')
plt.show();

Observation

1. Maruti has most no of sales followed by hyundai and honda.
2. Whereas force, Rolls-royce, ferrari and Mercedes_AMG have least no. of sales.
3. Hyundai i20 is highest selling car followed by Swift dezire and swift.
4. Whereas force Gurkha, Hyundai aura, ferrari GTC4Lusso ,etc are least selling cars.

3.2 Brands and Cars wrt Total Selling Price

### Brands wrt Total Selling Price
pd.DataFrame(dataset.groupby('brand')['selling_price'].sum().sort_values(ascending=False)).rename(columns={'selling_price':'Total Selling Price'})

	Total Selling Price
brand
Maruti	2402298875
Hyundai	1701766000
BMW	1171390000
Toyota	1082111000
Honda	911360000
Mercedes-Benz	823437000
Mahindra	787825000
Ford	491764000
Audi	374443000
Volkswagen	316529000
Tata	288200000
Skoda	261486000
Renault	233213000
Land Rover	192344000
Jaguar	153539000
Porsche	108385000
Volvo	74594000
Jeep	73628000
Kia	55528000
Datsun	54488000
Lexus	47365000
Ferrari	39500000
Mini	37105000
MG	33306000
Bentley	27800000
Rolls-Royce	24200000
Maserati	12200000
Nissan	10509000
Isuzu	9640000
Mercedes-AMG	5100000
ISUZU	3795000
Force	700000

### Top 10 brands
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('brand')['selling_price'].sum().sort_values(ascending=False)[:10],
            x=dataset.groupby('brand')['selling_price'].sum().sort_values(ascending=False).index[:10])
plt.ylabel('Total Sales in Rupees',fontsize=15, fontweight='bold')
plt.xlabel("Brand",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Top 10 brands wrt Total Selling price",fontsize=15, fontweight='bold')
plt.show();

### Bottom 10 brands
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('brand')['selling_price'].sum().sort_values(ascending=False)[-10:],
            x=dataset.groupby('brand')['selling_price'].sum().sort_values(ascending=False).index[-10:])
plt.ylabel('Total Sales in Rupees',fontsize=15, fontweight='bold')
plt.xlabel("Brand",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Bottom 10 brands wrt Total Selling price",fontsize=15, fontweight='bold')
plt.show();

### Top 10 Cars wrt Total Selling Price
pd.DataFrame(dataset.groupby('car_name')['selling_price'].sum().sort_values(ascending=False)[:10]).rename(columns={'selling_price':'Total Selling Price'})

	Total Selling Price
car_name
Toyota Innova	639831000
Hyundai i20	487790000
Honda City	468565000
Maruti Swift Dzire	459811000
Maruti Swift	365551000
Toyota Fortuner	364188000
Hyundai Creta	343551000
BMW 5	337804000
Mahindra XUV500	329429000
Hyundai Verna	318210000

### Bottom 10 Cars wrt Total Selling Price
pd.DataFrame(dataset.groupby('car_name')['selling_price'].sum().sort_values(ascending=False)[-10:]).rename(columns={'selling_price':'Toatl Selling Price'})

	Toatl Selling Price
car_name
Honda CR	4415000
ISUZU MUX	3795000
Datsun redi-GO	3212000
Isuzu MUX	2300000
Maruti Dzire ZXI	2200000
Nissan X-Trail	2135000
Hyundai Aura	900000
Maruti Dzire LXI	885000
Tata Altroz	730000
Force Gurkha	700000

### Top 10 selling cars 
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('car_name')['selling_price'].sum().sort_values(ascending=False)[:10],
            x=dataset.groupby('car_name')['selling_price'].sum().sort_values(ascending=False).index[:10])
plt.ylabel('Total Sales in Rupees',fontsize=15,  fontweight='bold')
plt.xlabel("Car",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Top 10 cars wrt total Selling price",fontsize=15, fontweight='bold')
plt.show();

### Bottom 10 selling cars 
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('car_name')['selling_price'].sum().sort_values(ascending=False)[-10:],
            x=dataset.groupby('car_name')['selling_price'].sum().sort_values(ascending=False).index[-10:])
plt.ylabel('Total Sales in Rupees',fontsize=15,  fontweight='bold')
plt.xlabel("Car",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Bottom 10 cars wrt Total Selling price",fontsize=15, fontweight='bold')
plt.show();

Observation

1. Maruti has highest Total amount of sales followed by hyundai and BMW.
2. Whereas force, ISUZU, Mercedes_AMG and Nissan have least Total amount of sales in that order.
3. Toyota Innova has highest total amount of sales followed by Hyundai i20 and Honda City.
4. Whereas force Gurkha, Tata Altroz, Maruti dzire LXI have least total amount of car sales.

3.3 Brands and Cars wrt Costliest Cars

### Brands wrt Costliest Car
pd.DataFrame(dataset.groupby('brand')['selling_price'].max().sort_values(ascending=False)).rename(columns={'selling_price':'Costliest Car Price'})

	Costliest Car Price
brand
Ferrari	39500000
Rolls-Royce	24200000
Bentley	14500000
Mercedes-Benz	13000000
Porsche	11100000
Land Rover	9200000
BMW	8500000
Volvo	8195000
Lexus	8000000
Audi	6800000
Jaguar	6300000
Maserati	6200000
Jeep	5600000
Mercedes-AMG	5100000
Mini	3875000
Toyota	3650000
Skoda	3550000
Kia	3525000
Ford	3200000
Honda	3200000
Mahindra	2950000
Hyundai	2600000
Isuzu	2300000
MG	2075000
ISUZU	1900000
Tata	1750000
Nissan	1450000
Volkswagen	1250000
Maruti	1225000
Renault	1155000
Force	700000
Datsun	650000

### Top 10 brand wrt Costliest car
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('brand').selling_price.max().sort_values(ascending=False)[:10],
            x=dataset.groupby('brand').selling_price.max().sort_values(ascending=False).index[:10])
plt.ylabel('Price of Costliest Car in Rupees',fontsize=15, fontweight='bold')
plt.xlabel("Brand",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Top 10 brands wrt Costliest Car",fontsize=15, fontweight='bold')
plt.show();

### Bottom 10 brand wrt Costliest cars 
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('brand').selling_price.max().sort_values(ascending=False)[-10:],
            x=dataset.groupby('brand').selling_price.max().sort_values(ascending=False).index[-10:])
plt.ylabel('Price of Costliest Car in Rupees',fontsize=15, fontweight='bold')
plt.xlabel("Brand",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Bottom 10 brands wrt Costliest Car",fontsize=15, fontweight='bold')
plt.show();

### Top 10 Costliest Car
pd.DataFrame(dataset.groupby('car_name')['selling_price'].max().sort_values(ascending=False)[:10]).rename(columns={'selling_price':'Car Price'})

	Car Price
car_name
Ferrari GTC4Lusso	39500000
Rolls-Royce Ghost	24200000
Bentley Continental	14500000
Mercedes-Benz S-Class	13000000
Porsche Cayenne	11100000
Land Rover Rover	9200000
BMW 7	8500000
BMW Z4	8250000
Volvo XC	8195000
BMW X5	8100000

### Bottom 10 Costliest Car
pd.DataFrame(dataset.groupby('car_name')['selling_price'].max().sort_values(ascending=False)[-10:]).rename(columns={'selling_price':'Car Price'})

	Car Price
car_name
Maruti Wagon R	625000
Maruti Celerio	595000
Maruti S-Presso	550000
Renault KWID	550000
Maruti Dzire LXI	500000
Hyundai i10	500000
Maruti Eeco	490000
Maruti Alto	485000
Datsun redi-GO	435000
Datsun RediGO	425000

### Top 10 costliest cars
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('car_name').selling_price.max().sort_values(ascending=False)[:10],
            x=dataset.groupby('car_name').selling_price.max().sort_values(ascending=False).index[:10:])
plt.ylabel('Price of Car in Rupees',fontsize=15, fontweight='bold')
plt.xlabel("Car",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Top 10 Costliest Cars",fontsize=15, fontweight='bold')
plt.show();

### Bottom 10 costliest cars
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('car_name').selling_price.max().sort_values(ascending=False)[-10:],
            x=dataset.groupby('car_name').selling_price.max().sort_values(ascending=False).index[-10:])
plt.ylabel('Price of Car in Rupees',fontsize=15, fontweight='bold')
plt.xlabel("Car",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Bottom 10 Costliest Cars",fontsize=15, fontweight='bold')
plt.show();

Observation

1. Ferrari has Costliest car followed by Rolls-Royce and Bentley.
2. Datsun has cheapest car followed by Force and Renault.
3. Ferrari GTC4Lusso is costliest car followed by Rolls-Royce Ghost and bentley Continental.
4. Datsun RediGO is cheapest car followed by Maruti Alto and Maruti Eeco.

3.4 Brands and Cars wrt Mileage

### Brand wrt Average Mileage
pd.DataFrame(dataset.groupby('brand')['mileage'].mean().sort_values(ascending=False)).rename(columns={'mileage':'Average mileage'})

	Average mileage
brand
Maruti	22.425520
Renault	22.073586
Datsun	21.215647
Lexus	20.676667
Ford	19.970335
Honda	19.901463
Maserati	19.820000
Tata	19.780855
Hyundai	19.590650
Volkswagen	18.690977
Mini	18.287647
Skoda	17.663904
BMW	17.450046
Kia	17.323125
Force	17.000000
Jeep	16.236585
Nissan	16.151818
Audi	16.119737
Jaguar	16.085172
Mahindra	15.864314
MG	15.620526
Volvo	14.860000
ISUZU	13.800000
Mercedes-Benz	13.531054
Porsche	13.515714
Toyota	13.207414
Land Rover	13.038400
Isuzu	12.600000
Mercedes-AMG	11.900000
Rolls-Royce	10.200000
Bentley	8.033333
Ferrari	4.000000

### Top 10 brand wrt Mileage of cars 
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('brand').mileage.mean().sort_values(ascending=False)[:10],
            x=dataset.groupby('brand').mileage.mean().sort_values(ascending=False).index[:10])
plt.ylabel('Mileage of Cars in KM/Lt',fontsize=15, fontweight='bold')
plt.xlabel("Brand",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Top 10 brands wrt Mileage of Car",fontsize=15, fontweight='bold')
plt.show();

### Top 10 brand wrt Mileage of cars 
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('brand').mileage.mean().sort_values(ascending=False)[-10:],
            x=dataset.groupby('brand').mileage.mean().sort_values(ascending=False).index[-10:])
plt.ylabel('Mileage of Cars in KM/Lt',fontsize=15, fontweight='bold')
plt.xlabel("Brand",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Bottom 10 brands wrt Mileage of Car",fontsize=15, fontweight='bold')
plt.show();

### Top 10 car wrt Mileage
pd.DataFrame(dataset.groupby('brand')['mileage'].mean().sort_values(ascending=False)[:10]).rename(columns={'mileage':'Average mileage'})

	Average mileage
brand
Maruti	22.425520
Renault	22.073586
Datsun	21.215647
Lexus	20.676667
Ford	19.970335
Honda	19.901463
Maserati	19.820000
Tata	19.780855
Hyundai	19.590650
Volkswagen	18.690977

### Bottom 10 car wrt Mileage
pd.DataFrame(dataset.groupby('brand')['mileage'].mean().sort_values(ascending=False)[-10:]).rename(columns={'mileage':'Average mileage'})

	Average mileage
brand
ISUZU	13.800000
Mercedes-Benz	13.531054
Porsche	13.515714
Toyota	13.207414
Land Rover	13.038400
Isuzu	12.600000
Mercedes-AMG	11.900000
Rolls-Royce	10.200000
Bentley	8.033333
Ferrari	4.000000

### Top 10 car wrt Mileage 
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('car_name').mileage.mean().sort_values(ascending=False)[:10],
            x=dataset.groupby('car_name').mileage.mean().sort_values(ascending=False).index[:10])
plt.ylabel('Mileage of Cars in KM/Lt',fontsize=15, fontweight='bold')
plt.xlabel("Car",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Top 10 Car wrt Mileage",fontsize=15, fontweight='bold')
plt.show();

### bottom 10 car wrt Mileage 
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('car_name').mileage.mean().sort_values(ascending=False)[-10:],
            x=dataset.groupby('car_name').mileage.mean().sort_values(ascending=False).index[-10:])
plt.ylabel('Mileage of Cars in KM/Lt',fontsize=15, fontweight='bold')
plt.xlabel("Car",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Bottom 10 Car wrt Mileage",fontsize=15, fontweight='bold')
plt.show();

Observation

1. Maruti cars have highest average mileage followed by Renault and Datsun.
2. Ferrari cars have least average Mileage followed by Bentely and Rolls Royce.
3. Tata Tiago has highest average mileage followed by Maruti Ciaz and Maruti Vitara.
4. Ferrari GTC4Lusso has least average Mileage followed by bentley Continental and Jeep Wrangler.

3.5 Brands and Cars wrt Max Power

### Brands wrt Mean Max Power
pd.DataFrame(dataset.groupby('brand')['max_power'].mean().sort_values(ascending=False)).rename(columns={'max_power':'Mean Max Power'})

	Mean Max Power
brand
Ferrari	601.000000
Bentley	592.666667
Rolls-Royce	563.000000
Mercedes-AMG	362.070000
Porsche	299.028571
Maserati	270.940000
Lexus	219.912222
Mercedes-Benz	210.780120
Land Rover	210.237400
BMW	207.604266
Jaguar	205.987414
Volvo	205.350000
Audi	181.430158
ISUZU	175.000000
Jeep	172.375610
Mini	151.167059
MG	150.829474
Isuzu	139.795714
Toyota	135.806793
Kia	130.025000
Skoda	126.194414
Nissan	122.119091
Tata	109.823159
Mahindra	109.177568
Honda	104.340732
Ford	95.931198
Volkswagen	92.629837
Hyundai	92.459993
Force	80.800000
Renault	75.055579
Maruti	74.530935
Datsun	62.645294

### Top 10 Brands wrt Mean Max Power 
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('brand').max_power.mean().sort_values(ascending=False)[:10],
            x=dataset.groupby('brand').max_power.mean().sort_values(ascending=False).index[:10])
plt.ylabel('Mean max Power of Cars',fontsize=15, fontweight='bold')
plt.xlabel("Brand",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Top 10 Brand wrt Mean Max Power",fontsize=15, fontweight='bold')
plt.show();

### Bottom 10 Brands wrt Mean Max Power 
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('brand').max_power.mean().sort_values(ascending=False)[-10:],
            x=dataset.groupby('brand').max_power.mean().sort_values(ascending=False).index[-10:])
plt.ylabel('Mean max Power of Cars',fontsize=15, fontweight='bold')
plt.xlabel("Brand",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Bottom 10 Brand wrt Mean Max Power",fontsize=15, fontweight='bold')
plt.show();

### Top 10 Cars wrt Max Power
pd.DataFrame(dataset.groupby('car_name')['max_power'].max().sort_values(ascending=False)[:10]).rename(columns={'max_power':'Max Power'})

	Max Power
car_name
Bentley Continental	626.00
Ferrari GTC4Lusso	601.00
Rolls-Royce Ghost	563.00
Porsche Cayenne	500.00
Mercedes-Benz S-Class	459.00
BMW 6	450.00
BMW 7	402.00
Porsche Panamera	394.30
Mercedes-AMG C	362.07
Mercedes-Benz GL-Class	362.00

### Bottom 10 Cars wrt Max Power
pd.DataFrame(dataset.groupby('car_name')['max_power'].max().sort_values(ascending=False)[-10:]).rename(columns={'max_power':'Max Power'})

	Max Power
car_name
Datsun GO	76.43
Maruti Eeco	73.00
Renault Triber	72.00
Hyundai Santro	68.07
Datsun RediGO	68.00
Maruti Alto	67.10
Maruti Celerio	67.05
Datsun redi-GO	67.05
Renault KWID	67.00
Maruti S-Presso	67.00

### Top 10 car wrt Max Power 
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('car_name').max_power.max().sort_values(ascending=False)[:10],
            x=dataset.groupby('car_name').max_power.max().sort_values(ascending=False).index[:10])
plt.ylabel('max Power of Cars',fontsize=15, fontweight='bold')
plt.xlabel("Car",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Top 10 Car wrt Max Power",fontsize=15, fontweight='bold')
plt.show();

### Top 10 car wrt Max Power 
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('car_name').max_power.max().sort_values(ascending=False)[-10:],
            x=dataset.groupby('car_name').max_power.max().sort_values(ascending=False).index[-10:])
plt.ylabel('max Power of Cars',fontsize=15, fontweight='bold')
plt.xlabel("Car",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Bottom 10 Car wrt Max Power",fontsize=15, fontweight='bold')
plt.show();

Observation

1. Ferrari has highest Mean Max Power followed by Bentley and Rolls-Royce.
2. Datsun has least Mean Max Power followed by Maruti and Renault.
3. Bentley Continental has highest Mean Max Power followed by Ferrari GTC4lusso and Rolls-Royce Ghost.
4. Maruti S-Presso least Mean Max Power followed by Renault KWID and Datsun RediGo.

3.6 Brands and Cars wrt Km Driven

### Brands wrt Mean KM Driven
pd.DataFrame(dataset.groupby('brand')['km_driven'].mean().sort_values(ascending=False)).rename(columns={'km_driven':'Mean KM Driven'})

	Mean KM Driven
brand
Toyota	91840.925222
Mahindra	73075.585586
Isuzu	69876.000000
Land Rover	65874.760000
Skoda	64784.492492
Volkswagen	62849.574919
Volvo	61232.300000
Force	60000.000000
Mercedes-Benz	59471.605422
Audi	58600.231579
Porsche	58356.952381
Ford	58324.350515
ISUZU	56014.500000
Honda	53525.912602
Maruti	51386.572268
BMW	51094.483945
Hyundai	50346.743225
Renault	47803.174573
Tata	47466.076010
Jaguar	44499.103448
Nissan	37853.363636
Jeep	33401.365854
Datsun	33003.576471
Mini	32210.705882
Lexus	28210.444444
Bentley	25500.000000
Mercedes-AMG	24000.000000
Maserati	12250.000000
MG	11538.684211
Kia	9255.437500
Rolls-Royce	5000.000000
Ferrari	3800.000000

### Top 10 Brands wrt Mean Km Driven
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('brand').km_driven.mean().sort_values(ascending=False)[:10],
            x=dataset.groupby('brand').km_driven.mean().sort_values(ascending=False).index[:10])
plt.ylabel('Mean Km Driven in KM',fontsize=15, fontweight='bold')
plt.xlabel("Brand",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Top 10 Brand wrt Mean Km Driven",fontsize=15, fontweight='bold')
plt.show();

### Bottom 10 Brands wrt Mean Km Driven
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('brand').km_driven.mean().sort_values(ascending=False)[-10:],
            x=dataset.groupby('brand').km_driven.mean().sort_values(ascending=False).index[-10:])
plt.ylabel('Mean Km Driven in KM',fontsize=15, fontweight='bold')
plt.xlabel("Brand",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Bottom 10 Brand wrt Mean Km Driven",fontsize=15, fontweight='bold')
plt.show();

### Top 10 car wrt Km Driven
pd.DataFrame(dataset.groupby('car_name')['km_driven'].max().sort_values(ascending=False)[10:]).rename(columns={'km_driven':'Max KM Driven'})

	Max KM Driven
car_name
Ford Figo	570000
Honda Jazz	525000
Mahindra Bolero	500000
Maruti Ciaz	480000
Maruti Vitara	480000
...	...
Maserati Quattroporte	9500
Rolls-Royce Ghost	5000
Hyundai Aura	4500
Tata Altroz	3800
Ferrari GTC4Lusso	3800

111 rows × 1 columns

### Bottom 10 car wrt Km Driven
pd.DataFrame(dataset.groupby('car_name')['km_driven'].max().sort_values(ascending=False)[-10:]).rename(columns={'km_driven':'Max KM Driven'})

	Max KM Driven
car_name
Maruti XL6	23000
Maserati Ghibli	15000
Kia Carnival	14000
BMW X4	14000
Maruti S-Presso	12000
Maserati Quattroporte	9500
Rolls-Royce Ghost	5000
Hyundai Aura	4500
Tata Altroz	3800
Ferrari GTC4Lusso	3800

### Top 10 car wrt Km Driven
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('car_name').km_driven.max().sort_values(ascending=False)[:10],
            x=dataset.groupby('car_name').km_driven.max().sort_values(ascending=False).index[:10])
plt.ylabel('Km Driven of Cars',fontsize=15, fontweight='bold')
plt.xlabel("Car",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Top 10 Car wrt Km Driven",fontsize=15, fontweight='bold')
plt.show();

### Bottom 10 car wrt Km Driven
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('car_name').km_driven.max().sort_values(ascending=False)[-10:],
            x=dataset.groupby('car_name').km_driven.max().sort_values(ascending=False).index[-10:])
plt.ylabel('Km Driven of Cars',fontsize=15, fontweight='bold')
plt.xlabel("Car",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Bottom 10 Car wrt Km Driven",fontsize=15, fontweight='bold')
plt.show();

Observation

1. Toyota cars have highest Mean KM Driven followed by Mahindra and isuzu.
2. Ferrari cars have least Mean KM Driven followed by Rolls-Royce and Kia.
3. Mahindra XUV500 has highest KM Driven followed by Mercedes-Benz E-class and Toyota Innova.
4. Ferrari GTC4lusso has least KM Driven followed by Tata Altroz and Hyundai Aura.

3.7 KM Driven and Seller Type vs Selling Price

### Km driven wrt selling price
plt.figure(figsize=(15,8))
sns.scatterplot(y=dataset['selling_price'], x=dataset['km_driven'], hue=dataset['seller_type'])
plt.ylabel('Selling Price of Cars in Rupees',fontsize=15, fontweight='bold')
plt.xlabel("Km Driven in KM",fontsize=15, fontweight='bold')
plt.xlim(0,800000) #limit is used for scaling the axis
plt.ylim(0,10000000)
plt.title("KM Driven and Seller Type vs Selling Price",fontsize=15, fontweight='bold')
plt.show();

Observation

1. It is difficult to sell car if Km Driven is greater than 20K.
2. As the Km driven increases the the Selling price of cars Decreases.
3. It is easy to sell cars at higher price if it is sold by a dealer.

3.8 KM Driven and transmission_type vs Selling Price

### Km driven wrt selling price
plt.figure(figsize=(15,8))
sns.scatterplot(y=dataset['selling_price'], x=dataset['km_driven'], hue=dataset['transmission_type'])
plt.ylabel('Selling Price of Cars in Rupees',fontsize=15, fontweight='bold')
plt.xlabel("Km Driven in KM",fontsize=15, fontweight='bold')
plt.xlim(0,800000) #limit is used for scaling the axis
plt.ylim(0,10000000)
plt.title("KM Driven and Transmission_type vs Selling Price",fontsize=15, fontweight='bold')
plt.show();

Observation

1. It is difficult to sell car if Km Driven is greater than 20K.
2. As the Km driven increases the the Selling price of cars Decreases.
3. Automatic Transmission cars sell at higher rate than manual Transmission cars.

3.9 KM Driven and fuel_type vs Selling Price

### Km driven wrt selling price
plt.figure(figsize=(15,8))
sns.scatterplot(y=dataset['selling_price'], x=dataset['km_driven'], hue=dataset['fuel_type'])
plt.ylabel('Selling Price of Cars in Rupees',fontsize=15, fontweight='bold')
plt.xlabel("Km Driven in KM",fontsize=15, fontweight='bold')
plt.xlim(0,800000) #limit is used for scaling the axis
plt.ylim(0,10000000)
plt.title("KM Driven and fuel_type vs Selling Price",fontsize=15, fontweight='bold')
plt.show();

Observation

1. It is difficult to sell car if Km Driven is greater than 20K.
2. As the Km driven increases the the Selling price of cars Decreases.
3. Majority of cars are either diesel or petrol.

3.10 KM Driven and seats vs Selling Price

### Km driven wrt selling price
plt.figure(figsize=(15,6))
sns.scatterplot(y=dataset['selling_price'], x=dataset['km_driven'], hue=dataset['seats'])
plt.ylabel('Selling Price of Cars in Rupees',fontsize=15, fontweight='bold')
plt.xlabel("Km Driven in KM",fontsize=15, fontweight='bold')
plt.xlim(0,800000) #limit is used for scaling the axis
plt.ylim(0,10000000)
plt.title("KM Driven vs Selling Price",fontsize=15, fontweight='bold')
plt.show();

Observation

1. It is difficult to sell car if Km Driven is greater than 20K.
2. As the Km driven increases the the Selling price of cars Decreases.
3. Market is dominated by 3, 4 and 6 seater cars.
4. As no. of seats decreases the selling price of cars increases.

3.11 Numerical features and categorical features vs Selling Price

for i in [feature for feature in numerical_features if feature not in ['selling_price', 'km_driven']]:
    for j in ['seller_type', 'fuel_type', 'transmission_type']:
        plt.figure(figsize=(15,6))
        sns.scatterplot(y=dataset['selling_price'], x=dataset[i], hue=dataset[j])
        plt.ylabel('Selling Price of Cars in Rupees',fontsize=15, fontweight='bold')
        plt.xlabel("{}".format(i),fontsize=15, fontweight='bold')
        plt.ylim(0,10000000)
        plt.title("{} and {} vs Selling Price".format(i,j),fontsize=15, fontweight='bold')
        plt.show();

Observation

1. As the vehicle age increase the selling price of vehicle decreases.
2. Dealers are able to sell the cars at higher price than individual owners.
3. Between 1 to 7 vehicle age seller has better chance to get a good price of car.
4. On an average disesl cars are more in demand than petrol car and they sell for higher price.
5. Majority of cars are either diesel or petrol.
6. Automatic transmission cars recieve higher selling price than manual transmission cars.
7. Majority of cars have mileage between 10 to 26 Km/Lt.
8. Cars having mileage between 10 to 20 may get higher selling price offer.
9. It is evident that CNG cars have more mileage than petrol and diesel.
10. Manual transmission cars have more Mileage in comparision to Automatic transmiossion cars.
11. Engine with 1000 to 3000 bhp get more buyer attention.
12. As the max power increases the selling price of cars also increases.
13. Majority of high power cars is sold by dealers.
14. Diesel cars have more power in comparision to Petrol cars.
15. Cars with high power generally comes with automatic transmission.
16. Majority of cars have 4,5 and 7 seats.
17. More than 6 seater car is usually diesel car.

3.12 Fuel Type analysis

### most cars sold on fuel type
plt.figure(figsize=(10,6))
sns.countplot(x=dataset['fuel_type'])
plt.ylabel("Count of Cars sold",fontsize=15, fontweight='bold')
plt.xlabel("Fuel Type",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Most Sold Car based on Fuel Type",fontsize=15, fontweight='bold')
plt.show();

Observation

1. Petrol varient have highest Number of cars followed by diesel and CNG.
2. Electric cars, CNG and LPG cars are very few in number.

### fuel type vs total selling price
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('fuel_type')['selling_price'].sum().sort_values(ascending=False),
            x=dataset.groupby('fuel_type')['selling_price'].sum().sort_values(ascending=False).index)
plt.ylabel('Total Selling Price in Rupees',fontsize=15, fontweight='bold')
plt.xlabel("Fuel Type",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Fuel Type vs Total Selling Price",fontsize=15, fontweight='bold')
plt.show();

Observation

1. Diesel cars have highest total sales followed by petrol.
2. The combined total sales of CNG, LPG and Electric is negligible as compared to Diesel and Petrol.

### fuel type vs average selling price
plt.figure(figsize=(10,6))
sns.barplot(y=dataset.groupby('fuel_type')['selling_price'].mean().sort_values(ascending=False),
            x=dataset.groupby('fuel_type')['selling_price'].mean().sort_values(ascending=False).index)
plt.ylabel('Mean of Selling Price in Rupees',fontsize=15, fontweight='bold')
plt.xlabel("Fuel Type",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Fuel Type vs Mean Selling Price",fontsize=15, fontweight='bold')
plt.show();

Observation

1. Electric cars have highest average selling price followed by Diesel and Petrol.
2. LPG has least average selling price.
3. Electric vehicles are twice costlier than diesel car and thrice costlier than petrol cars.

### Fuel Type vs Average Mileage
plt.figure(figsize=(10,6))
sns.barplot(y=dataset['mileage'],
            x=dataset['fuel_type'])
plt.ylabel('Mean of Mileage in KM/Lt',fontsize=15, fontweight='bold')
plt.xlabel("Fuel Type",fontsize=15, fontweight='bold')
plt.xticks(rotation=45)
plt.title("Fuel Type vs Mean Mileage",fontsize=15, fontweight='bold')
plt.show();

Observation

1. CNG cars have highest average Mileage followed by diesel, petrol and electric cars.
2. LPG has least average Mileage.

3.13 Vehicle age vs Selling Price

plt.subplots(figsize=(20,8))
sns.lineplot(x=dataset['vehicle_age'],y=dataset['selling_price'], color='green')
plt.ylabel('Selling_price', fontsize=15, fontweight='bold')
plt.xlabel('vehicle_age', fontsize=15, fontweight='bold')
plt.title('Vehicle Age vs Selling Price', fontsize=15, fontweight='bold')
plt.ylim(0,4000000)
plt.show()

Observation

1. As the vehicle age increases the selling price of vehicles decreases.
2. The exception is for cars having age between 22 to 25 years.

3.14 Vehicle age vs Mileage

plt.subplots(figsize=(20,8))
sns.lineplot(x=dataset['vehicle_age'],y=dataset['mileage'], color='green')
plt.ylabel('Mileage', fontsize=15, fontweight='bold')
plt.xlabel('vehicle_age', fontsize=15, fontweight='bold')
plt.title('Vehicle Age vs Mileage', fontsize=15, fontweight='bold')
plt.show()

Observation

1. As the vehicle age increases the Average mileage decreases.

3.15 Distribution of Numerical Features

### Checking distribution of numerical features

for feature in numerical_features:
    plt.figure(figsize=(15,6))
    plt.subplot(121)
    sns.histplot(data=dataset, x=feature, kde=True, bins=30)
    plt.title("{}'s distribution".format(feature),fontweight="bold", fontsize=15)
    
    plt.subplot(122)
    stats.probplot(dataset[feature], dist='norm', plot=plt)
    plt.title("{}'s Q-Q Plot".format(feature),fontweight="bold", fontsize=15)
    plt.show();

Observation

1. Vehicle age, Km Driven, Engine, max_power and selling price have right skewed distribution.
2. Mileage has close to normal distribution just a little bit right skewness is present.
3. Seat is a discreate feature.

3.16 Outliers in Numerical Features

### Checking outliers in numerical features

plt.figure(figsize=(20,25))
for feature in enumerate(numerical_features):
    plt.subplot(4, 2, feature[0]+1)
    sns.set(rc={'figure.figsize':(10,6)})
    sns.boxplot(data=dataset, x=feature[1], color='pink')
    plt.title("{}".format(feature[1]), fontweight="bold", fontsize=15)

Observation

1. Km driven, Max Power and Selling price has large no. of outliers on upper limit side.
2. Vehicle age and engine have less outliers as compared to above features.
3. Mileage has outliers on both sides.

3.17 Regplot for Numerical features vs Dependent feature

### plotting regplot for features vs dependent feature
plt.figure(figsize=(20,25))
for feature in enumerate([feature for feature in numerical_features if feature not in ['selling_price']]):
    plt.subplot(3, 2, feature[0]+1)
    sns.set(rc={'figure.figsize':(8,8)})
    sns.regplot(data=dataset, x=feature[1], y='selling_price')
    plt.xlabel(feature[1])
    plt.ylabel("Selling Price")
    plt.title("{} Vs Selling Price".format(feature[1]), fontweight='bold', fontsize=15)

3.18 Correlation between Numerical Features and Dependent Feature

corr=round(dataset[numerical_features].corr(),2)
corr

	vehicle_age	km_driven	mileage	engine	max_power	seats	selling_price
vehicle_age	1.00	0.33	-0.26	0.10	0.00	0.03	-0.24
km_driven	0.33	1.00	-0.10	0.19	0.04	0.19	-0.08
mileage	-0.26	-0.10	1.00	-0.63	-0.53	-0.44	-0.31
engine	0.10	0.19	-0.63	1.00	0.81	0.55	0.59
max_power	0.00	0.04	-0.53	0.81	1.00	0.17	0.75
seats	0.03	0.19	-0.44	0.55	0.17	1.00	0.12
selling_price	-0.24	-0.08	-0.31	0.59	0.75	0.12	1.00

### Plotting heatmap for visualising the correlation between features
sns.set(rc={'figure.figsize':(12,7)})
sns.heatmap(data=corr, annot=True,  vmin=-1, vmax=1, cmap="YlGnBu")

<AxesSubplot:>

Observation

1. Selling Price has negative correlation with vehicle age, Km driven, and Mileage.
2. Engine and Max power have moderate correlation with Selling Price.
3. Max Power and engine have high correlation.
4. Km Driven, Vehicle age and Seats have very less correlation with selling price.